Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Automated Theorem Proving on Metamath (val)
Loading...
56.5
Performance
700m policy+value a = 32
19.7464
29.2882
38.83
48.3718
Sep 7, 2020
Performance
Gain
Cumulative Success
Pass@8
Pass@32
Updated 4d ago
Evaluation Results
Method
Method
Links
Performance
Gain
Cumulative Success
Pass@8
Pass@32
700m policy+value a = 32
Parameters=700m, Test-...
2020.09
56.5
9.2
-
-
-
700m policy+value
Parameters=700m, Optim...
2020.09
47.21
4.6
-
-
-
700m WebMath
Parameters=700m, Pre-t...
2020.09
42.56
10.9
-
-
-
700m
Parameters=700m
2020.09
31.58
2.5
-
-
-
160m
Parameters=160m, Archi...
2020.09
28.96
7.8
-
-
-
MetaGen-IL
Description=Baseline a...
2020.09
21.16
-
-
-
-
Supervised
training=supervised
2022.05
-
-
-
61
65.4
Evariste
training=online training
2022.05
-
-
82.6
81
81.2
Feedback
Search any
task
Search any
task